On September 22, grocer Trader Joe’s announced a recall of 16oz. plastic jars of Creamy Salted Valencia Peanut Butter. The original recall announcement, posted on the 22nd, contained the following language:

Out of the utmost caution and care for our customers, Trader Joe’s is voluntarily recalling its Creamy Salted Valencia Peanut Butter because of potential contamination with Salmonella. At this time, no confirmed illnesses directly linked to this product have been reported to Trader Joe’s.

While we have no confirmed information that suggests this peanut butter is unsafe to eat, there is nothing more important to us than the health and safety of our customers and crew, and the quality of our products.

On September 24, Drew Falkenstein, a Seattle-based food safety attorney and fantastic food safety blogger, took issue with the language, noting that on the same day of the recall, the CDC had announced that 14 of the 28 cases of a Salmonella Bredeney cluster had been interviewed, with 14/14 cases reporting shopping at Trader Joe’s, and 12/14 cases reporting consumption of Trader Joe’s Creamy Salted Valencia Peanut Butter in the week before getting sick. Wrote Falkenstein:

That is called playing fast and loose with the facts. I won’t waste time speculating why Trader Joes denies the link between its peanut butter and this salmonella outbreak. Any observer of outbreaks and recalls knows that it is something like, “no product has tested positive, so it could not have been us.”

The more prudent course would have been to acknowledge the problem, which is clearly based on solid epidemiological proof, and focus on why the problem occurred. In its September 22 statement, the CDC states “Fourteen (100%) of 14 ill persons interviewed report shopping at Trader Joe’s locations across the United States. Twelve (86%) of 14 ill persons report eating Trader Joe’s Valencia Creamy Salted Peanut Butter made with Sea Salt in the week before becoming ill.” Recall notices are really not the right forum to conduct public relations.

As of 10:30pm CST on September 24th, without any additional updates from the CDC having been released to the public, Trader Joe’s recall announcement had been revised on the FDA’s website, without the language Falkenstein interpreted as a denial of association with Salmonella illnesses:

Out of the utmost caution and care for our customers, Trader Joe’s is voluntarily recalling its Creamy Salted Valencia Peanut Butter (sku 97111) because of potential contamination with Salmonella…At this time the Food and Drug Administration (FDA), Centers for Disease Control and Prevention, and state health agencies are investigating whether there are direct links between this product and reported illnesses. Trader Joe’s is fully cooperating with these authorities because there is nothing more important to us than the health and safety of our customers and crew, and the quality of our products.

This case study in public relations during food recall announcements brings to light an extremely important topic for food producers and distributors. What is the definition of “confirmed information?”

The Food Industry as a Partner in Outbreak Investigations

It is apparent that Trader Joe’s did not publicly consider an epidemiological hypothesis to be “confirmed information” at the time of the recall. Without specific information about the studies used to produce the epidemiological hypothesis, it is inappropriate to judge this decision, even considering the apparent strength of the hypothesis. We should remember that while it is often only an epidemiological hypothesis that determines the liability of the food industry during outbreaks, hypotheses can be wrong. For this reason, industry participation in outbreak investigations is absolutely crucial. Often, an industrial traceback of products associated with a cluster of illnesses can confirm or deny the plausibility of an epidemiological hypothesis. In some instances, an industry-led traceback may be equally or more efficient than a public health-led traceback.

Consider the high profile 2008 outbreak of Salmonella Saintpaul that was initially associated with tomato consumption before being associated with Jalapeno and Serrano peppers. No industry traceback was able to explain how tomatoes sent to so many states and restaurants could have become contaminated with the same strain of Salmonella Saintpaul, and no recalls seemed to be able to end the spread of the outbreak, which should have been over quickly due to the short shelf life of fresh tomatoes. Once epidemiologists identified an association with peppers, however, an industry traceback quickly led to a single distributor, and the outbreak strain of Salmonella was quickly identified on a farm that supplied the distributor with Jalapeno and Serrano peppers. The epidemiological hypothesis and subsequent removal of certain tomatoes from store shelves and restaurants was a devastating blow to the tomato industry, and in retrospect, was in all likelihood attributable to public health error.

In this case, did Trader Joe’s decide to announce a recall before considering the available purchase dates and/or lot codes to investigate the plausibility of the epidemiological hypothesis? Regardless of the choice of words in the recall announcement, a recall at this point would probably be considered relatively proactive and prudent in the interest of public health.

What Does the Food Industry Consider “Confirmed Information?”

Did Trader Joe’s consider internal traceback information “confirmed information,” or, as Falkenstein suggests, did Trader Joe’s consider an isolation of the outbreak strain of Salmonella from the actual end product or production plant “confirmed information?” At the same time, is it ever a good public relations strategy to suggest doubt in an epidemiological hypothesis in a recall announcement, or is that doubt better left unsaid?

These critical decisions are for senior risk managers and should probably be considered by all food producers and distributors before they are implicated in a recall. In the interest of preventing future illnesses, deaths and potential liability, it is probably prudent for food producers and distributors to issue recalls as soon as possible after being contacted by public health authorities; however, it can be a difficult decision for a company to recall food products based solely on a hypothesis from government agencies that have apparently been wrong before.

This case reminds us that proactive leadership between public health and the food industry is critical, and it is in the best interests of the food industry to train senior risk managers in epidemiological methods and inference before foodborne disease outbreaks occur.

Editor’s note: Drew Falkenstein, whose blog entry is mentioned above, is an attorney at Marler Clark law firm, which underwrites Food Safety News.